1

Remote Sensing, GIS and DEMApplications for Flood Monitoring

Prof. D. Nagesh KumarChairman, Centre for Earth Sciences

Professor, Dept. of Civil Engg.Indian Institute of Science

Bangalore – 560 012URL: http://www.civil.iisc.ernet.in/~nagesh

Outline

� Floods and Potential use of Remote Sensing

� Real Time Monitoring of Floods

� GIS for Flood Damage Assessment

� Digital Elevation Models (DEMs)

� Delineation of Flood-prone Areas using Modified Topographic Index for Mahanadi Basin

� Integrated Approach to Flood Management

� Conclusions

Introduction� Floods are the most common and widespread of all natural

disasters� Floods cause damage to houses, industries, public utilities

and property resulting in huge economic losses, apart from loss of lives

� Though it is not possible to control the flood disaster totally, by adopting suitable structural and non-structural measures the flood damages can be minimized

� For planning any flood management measure latest, reliable, accurate and timely information is required

� Remote sensing technology has made substantial contribution in every aspect of flood disaster management such as preparedness, prevention and relief.

Potential uses of Remote Sensing for Flood Management

� Flood inundation mapping and monitoring

� Rapid and scientific based damage assessment

� Monitoring and mapping of flood control works

� Monitoring and mapping of changes in the river course

� Identification of river bank erosion

� Identification of chronic flood prone areas

� Inputs for flood forecasting & warning models

Flood Inundation Mapping� Once a flood event occurs, information on flooded

areas and quick assessment of damages is required for planning flood relief activities

� Satellite remote sensing provides synoptic view of the flood-affected areas at frequent intervals for assessing • Progression and recession of the flood inundation in

short span of time which can be used for planning and organizing the relief operations effectively

Satellites and their Sensors used for Flood Mapping

2

IRS P6 AWiFS Image showing Various Flood Features IRS Satellite Image (FCC) showing

the Ganga River Basin

IRS Satellite Image (FCC) showing the Brahmaputra River Basin

Real-Time Monitoring of Floods� Whenever a flood event occurs in India, NRSC

provides flood maps showing flood affected areas and flood damage statistics in near real time

� Presently, this information is being provided to the user departments within 6 hours after procuring the satellite data

� During the last one and half decades, major flood events in Assam, Bihar, Uttar Pradesh, West Bengal, Orissa, Andhra Pradesh, Jammu & Kashmir, Rajasthan, Punjab have been successfully mapped using Indian Remote Sensing (IRS) and other satellites.

Microwave Remote Sensing for Flood Mapping

� In adverse cloud conditions optical datafrom most of the satellites will not be useful

� Microwave SAR (Synthetic ApertureRadar) data has all weather capability.Radarsat provides such data

Radarsat mosaic of 4th and 7th July 2003 showing the

Brahmaputra River flood affected areas in Assam

Assam State

3

Monitoring Floods

Flood Year (1993), TM 432Non Flood Year (1988), TM 432

Satellite images (Resourcesat AWiFS) showing the original course of the Kosi River, India, and that developed after the breach due to flood in 2008

Bhatt et al., 2010

IRS CARTOSAT and LISS IV images showing before and after a flood event in the Kosi River, India

Bhatt et al., 2010

Temporal progression of Kosi River flood inundation in India during August 20–29, 2008

Bhatt et al., 2010

Satellite images (RADARSAT and IRS P6) showing the changes observed in the Kosi River course, India during 1998–2008

Bhatt et al., 2010

RS & GIS for Flood Hazard Mapping & Planning

4

Generation of Flood Damage StatisticsVillages marooned as on 7th July 2003 in Marigaon district, Assam

Detailed view of marooned villages

Utilization of Flood Imagesfrom RS & GIS

Delineation of Flood-prone Areas using Modified Topographic Index

for Mahanadi Basin

� Preparing and maintaining an accurate flood map is a difficult task.

� Ease in availability of surface elevation data has resulted in DEM based models.

� A simple method for delineation of flood-prone areas, Modified topographic index, is applied for the Mahanadi Basin

23

Introduction

� Topographic index (Kirkby, 1975) is defined as

TI is the topographic index

ad is the drained area per unit contour length

tan β is the local slope.

24

Topographic Index

5

� The modified topographic index (Manfreda, 2008) is given by

TIm is the modified topographic index

n is an exponent ≤ 1.

25

Modified Topographic Index (MTI)

� It allows the delineation of the portion of the basin as exposed to flood inundation assuming that it is the area characterised by the modified topographic index exceeding a given threshold TIms

� The threshold will be estimated by using a flooding map of the basin, which is assumed to have correct representation of flooding and non-flooding areas

� Modified topographic index map is compared with flood inundation map, and value of modified topographic index above which area is considered as inundated is obtained

26

Delineation using MTI

� To estimate the threshold, two error functions are defined

� ER1 - Percentage of non-flooded area whose value of the modified topographic index is greater than the threshold

� ER2 - Percentage of flooded area which has a modified topographic index value less than the threshold. 27

Delineation using MTI (Contd.)� The objective is to define a threshold value which

minimizes both errors in the delineation of the flood inundation areas.

� The sum of two errors (ER1 + ER2) represents an objective function that can be used for the estimation of the two parameters TIms and n.

� An iterative algorithm is used on this function to search for a minimum value of (ER1 + ER2), to obtain the two parameters.

28

Delineation using MTI (Contd.)

� The study area lies between East longitudes 80°30’ and 86° 50’, and North latitudes 19° 15’ and 23° 35’.

� Length of the river is about 900 km, and it has a catchment area of approximately 1,41,600 km2.

� Climate in the basin of Mahanadi is predominantly sub-tropical.

� Annual rainfall varies from 1143 mm to 2032 mm over the entire basin, average being 1438.1 mm.

29

Study Area

Mahanadi Basin Map

30

6

� Digital Elevation Models (DEMs)

� Flood inundation maps for three flood events (2006, 2008 and 2011) were used.

� Annual maximum discharge data at Naraj for a period of 38 years was used for flood frequency analysis

31

Data Used

DEM Spatial Resolution (m)

ASTER Global DEM

30

SRTM 92

GMTED2010

228

455

911

32

ASTER Global DEM of Mahanadi Basin

33

Source: NRSC, Hyderabad

Flood inundation maps for 2006, 2008 and 2011

� GIS analysis was performed on the DEMs to obtain the specific catchment area maps and slope maps.

� Analysis was also performed on the flood inundation maps obtained from NRSC.

� ArcMap, ERDAS Imagine and MapWindow GIS were used to perform all the operations.

� Iterations are carried out for each value of the exponent to obtain modified topographic index.

� It is compared with the flood inundation map to obtain ‘n’ and ‘TIms’ having the minimum error.

34

Methodology

35

Flowchart to Obtain Minimum Error, 'n' and 'TIms'

� ‘n’ and ‘TIms’ obtained are used to produce flood inundation maps.

� Flood frequency analysis is performed by fitting the annual maximum flow data to three probability distributions: normal, log-normal and Gumbel distributions.

� Tests for goodness of fit were performed by using the χ2 test and the Kolmorgov-Smirnoff test.

36

Methodology (Contd.)

7

Event (Year)

Magnitude (1000 cumecs)

2006

(36.33)

2011

(38.71)

2008

(44.76)

DEM Resolution (m) n

900 0.01 0.01 0.01

450 0.04 0.01 0.01

225 0.04 0.01 0.01

90 0.03 0.01 0.01

30 0.01 0.01 0.01

DEM Resolution (m) TIms

900 6.13 5.49 5.27

450 6.24 5.21 5.08

225 5.61 4.78 4.09

90 5.28 4.38 4.32

30 3.22 3.22 2.87

DEM Resolution (m) ER1 + ER2

900 43.63 37.23 38.70

450 42.63 43.09 39.88

225 39.96 43.97 21.84

90 32.42 37.98 35.51

30 17.40 18.51 18.39

37

Results

Parameters Obtained from the Method and the Errors Corresponding to the Parameters for the Three Flood Events 

� The error reduces as the spatial resolution of the DEM reduces.

� It was also noted that ER1 (the over-estimation) was significantly larger than ER2 in all the cases.

� As the flood magnitude increases, TIms has reduced, indicating that a larger area will be under flood inundation.

38

Results (Contd.)

39

Flood Inundation Map for the 2006 Flood Event from SRTM DEM (90 m)

40

Flood Inundation Map for the 2008 Flood Event from SRTM DEM (90 m)

41

Flood Inundation Map for the 2011 Flood Event from SRTM DEM (90 m)

Integrated Approach to Flood Management

8

Integrated Flood Management (WMO, 2009)

Integrated Flood Management Plan

� Manage the water cycle as a whole

� Integrate land and water management

� Manage risk and uncertainty

� Adopt a best mix of strategies

� Ensure a participatory approach and

� Adopt integrated hazard management approaches.

Strategies and Options for Flood Management (WMO, 2009)

Integration of RS, GIS, DEM and Hydrological Models

RS GIS DEM Hydrological Model

Conclusions

� Strong potential for use of RS, GIS and DEM for Flood Hazard planning, mitigation and management

� Proper image processing of remotely sensed data, DEM and spatio-temporal analyses with GIS would be very effective for Flood Management

Thank you

9

Sources

� http://www.nrsc.gov.in/flood1.html� Bhanumurthy, V, Manjusree, P, Srinivasa Rao, G, (2010),

Flood Disaster Management, Chapter 12 In Remote Sensing Applications, Eds. PS Roy, RS Dwivedi and D Vijayan, National Remote Sensing Center, Hyderabad.

� WMO (2009), Integrated Flood Management – Concept Paper, WMO No. 1047, Associated Programme on Flood Management (APFM), World Meteorological Organization (WMO), Geneva, Switzerland.

� Manfreda, S., Sole, A., and Leo, M D. “Detection of flood prone areas using digital elevation models”. Journal of Hydrologic Engineering, ASCE, 16(10), 781-790, 2011

� Zhang, W., and Montgomery, D. R. “Digital elevation model grid size, landscape representation, and hydrologic simulations.” Water Resources Research, 30(4), 1019–1028, 1994.